[Martin Taylor 990114 14:22]
[From Rick Marken (990114.0750)]
The important point of the example (for me) was not the
accuracy of the belief inn term of physical law but the
"surprise" that occurs when what is actually perceived
differs from what one believes will be perceived. I
submit that the surprise reaction is evidence that there
is a reference for what will be perceived and that what is
perceived differs from this reference (there is an error).
The nature of control is that there is a reference value for a reference
signal, and that reference value may differ from a perceptual value,
causing output that influences the perceptual value. A difference between
the reference value and the perceptual value is normal, since the reason
for control is that external influences also affect the perceptual
value. If they didn't, there would no need to have a control system.
A simple outflow command system would do.
A "surprise reaction" is the last thing I would expect to accompany a
difference between a reference and the actual value of the perceptual
signal, given that _every_ elementary control unit in the whole
hierarchy experiences such differences much of the time. The normal
consequence is unsurprised action to reduce the difference.
It seems to me that surprise when perception differs from belief
is evidence __against__ the belief being a reference value for the
perception.
Martin